> ## Documentation Index
> Fetch the complete documentation index at: https://docs.biscuit.so/llms.txt
> Use this file to discover all available pages before exploring further.

# AI Models: Every model Biscuit supports

> The full list of AI models available in Biscuit, from Google, OpenAI, Anthropic and xAI

Biscuit supports AI models from Google, OpenAI, Anthropic, xAI and  We support open-weight models as well.

> [!NOTE]
>
> This page lists the most commonly used models. It isn't a complete list: Biscuit supports more models, and new ones are added as providers release them.

## Text models

### Anthropic

| Model | Description |
| - | - |
| Claude Fable 5 | Most capable Claude model |
| Claude Opus 5.5 | Long-running agentic and knowledge work |
| Claude Sonnet 5.5 | Fast and strong at everyday tasks |
| Claude Haiku 5.5 | Fastest and cheapest, for high-volume tasks |
| Claude Opus 5 | Strong reasoning |
| Claude Sonnet 5 | High performance for coding and agents |
| Claude 4 Opus | High quality |
| Claude 4 Sonnet | Balanced speed and quality |
| Claude 4 Haiku | Fast |

### Google

| Model | Description |
| - | - |
| Gemini 4 Argon | Flagship reasoning model for complex reasoning, math and code |
| Gemini 3.6 Flash | Newest fast model |
| Gemini 3.5 Flash | Stable fast model |
| Gemini 3 Pro | High quality |
| Gemini 3 Flash | Fast |
| Gemini 2.5 Pro | High quality |
| Gemini 2.5 Flash | Fast and efficient |

### OpenAI

| Model | Description |
| - | - |
| GPT-6.1 Sol | Strong all-rounder and agentic work |
| GPT-6.1 Terra | Fast mid-tier model, strong at screenshots, visual reasoning and writing |
| GPT-6 Sol | Strong all-rounder |
| GPT-6 Luna | Small and fast |
| GPT-5.5 | Advanced model |
| GPT-5.4 | Flagship of the GPT-5 generation |
| GPT-5.4 mini | Compact |
| GPT-5.4 nano | Very fast |
| GPT-5.2 | Balanced |
| GPT-5 mini | Compact |
| GPT-4.1 | General purpose |
| GPT-4.1 mini | Compact and fast |
| o4-mini | Reasoning |

## Image generation models

| Model | Provider | Description |
| - | - | - |
| GPT Image 2.5 Flare | OpenAI | Best quality, including transparent backgrounds |
| Gemini 2.1 Image (Nano Banana 2.1) | Google | Newest Gemini image model |
| Gemini 3.1 Flash Image (Nano Banana 2) | Google | Higher-quality fast images |
| Gemini 3.1 Flash Lite Image (Nano Banana 2 Lite) | Google | Fastest image generation |

## Web search models

| Model | Provider | Description |
| - | - | - |
| Gemini 3 Flash | Google | Fast web search |
| Grok 4.3 | xAI | Searches the web and posts on X |

## Voice models

| Model | Provider | Description |
| - | - | - |
| GPT Realtime 2 | OpenAI | Real-time voice conversations |

## FAQ

> [accordion] Do I need an API key to use these models?
>
> No. All models are built in and managed by Biscuit.

> [accordion] Which model does my app use by default?
>
> Biscuit picks the best model for each task, based on what you asked for and what your app needs.

> [accordion] Can I choose a specific model?
>
> Yes. Open your app, click **Edit**, and ask for the model by name, for example *"Use Claude Sonnet 5.5 in the summary workflow."*

> [accordion] Do AI providers train on my data?
>
> No. Biscuit's contracts with AI providers don't allow them to use your data to train or evaluate their models. See [AI security](https://docs.biscuit.so/ai-security.md).

> [accordion] Are new models added?
>
> Yes. Biscuit adds new models as providers release them.

## Up next

> [card] [AI security](https://docs.biscuit.so/ai-security.md)
>
> What Biscuit's AI can see, and how your data is handled by AI providers.

> [card] [IT and procurement FAQ](https://docs.biscuit.so/it-teams.md)
>
> Answers to common questions from IT and procurement teams.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.